We present the results of our institutional experience with FSRT for the treatment of skull base meningiomas. Our local control outcomes are consistent with other reports of 5 year progression free survival rates of 87–95.7% after treatment with either SRS [11–13, 15] or FSRT [8–11]. Since many of these patients are symptomatic at presentation (92% in our cohort) we feel that symptom outcome is a crucial end-point in addition to local control and overall survival, which tend to be universally excellent in published reports. This report adds to the existing literature on by providing a detailed analysis dissected by symptom domain in a large retrospective cohort. These patients were closely followed during RT, allowing detection of a high rate of symptom improvement early in the treatment course.
We found that in addition to providing excellent local control, FSRT resulted in durable symptom improvement of one or more symptoms in 57% of cases. For patients with reduced vision, FSRT durably improved or maintained vision in at least 71% of cases, demonstrating the utility of FSRT in function preservation. While less detailed, other series in the literature have also reported between a 5-46% rate of improvement in symptoms following radiation therapy with either SRS or FSRT [8, 9, 15, 16]. We have not attempted to compare these results with those reported from SRS or surgical series because =of multiple potential confounding factors such as case mix, recall bias and reporting bias.
We found a negative association between prior surgery, re-irradiation and increasing age with the likelihood of symptom improvement after FSRT. These finding suggest FSRT after prior definitive treatment is less likely to result in symptomatic improvements, particularly in older patients. Midline or bilateral location was a negative prognostic factor for symptom improvement, a finding which has not been reported previously. Midline and bilateral tumors were larger, received less radiation dose, and underwent resection prior to radiation more often. It is unclear whether this finding is the result of an intrinsic property of tumors of this region or a spurious finding as a result of multiple comparisons, and should be investigated in future studies. Of interest, despite evolving technology, the use of the BrainLab treatment planning system, a surrogate for patients treated prior to 2004 versus those treated after 2004, predict for either local control or symptom improvement.
When patient symptoms improved, they did so by the end of FSRT in one third of patients, and within two months of the end of FSRT in over half of patients. This observation of the early symptom improvement raises questions regarding the underlying mechanism. Such improvement was also reported in the literature to occur at 6 weeks after completion of IMRT  and FSRT . This time frame typically precedes radiographic shrinkage and likely occurs before the full tumoricidal effect of radiation. A placebo effect is unlikely to account for this symptoms relief, as patients were noted to have objective responses on cranial nerve examinations during weekly treatment visits. Although speculative, one possible explanation for restoration of cranial nerve function after radiation is re-distribution of vascular flow from tumor to the affected cranial nerve(s).
Conversely, while radiographic improvement does not necessarily correlate with symptomatic improvement, there was a strong relationship between radiographic tumor progression and symptom worsening. The mechanism of symptom worsening may differ from that of improvement. Indeed, the clinical finding of a new symptom may be an early indicator of radiographic progression.
Treatment was generally very well tolerated and adverse events were infrequent in our series, though estimates were conservative and included all possible treatment related events. Our reported rates of RION (2.2%) and radiation necrosis (2.2%) were low and may still over-estimate the actual adverse event rates, supporting FSRT as a safe treatment option.
Limitations to our study include limited documentation of symptom severity and relatively short median follow-up time. Although we recorded symptom severity, these were often recorded as subjective rather than objective measures. Some of patients did have objective measurement of symptom outcomes, such as with visual field testing, and these have typically correlated well with objective findings. We have chosen to not report the objective findings in this report as they represent only a subset of the overall cohort and may not be representative of overall outcomes. Additionally, given the difficulty in comparing severity between symptom domains, we did not use this measure in analysis.
The median follow-up time (4.4 years) is short given the range of patients who were treated from 1994 through 2009. Factors for this finding include losses to follow-up from patients who live far from our facility and to storage from paper records. Additionally, we strictly defined follow-up as the last evaluation by a radiation oncologist or a neurosurgeon, rather than the last evidence of patient being alive, in order to capture symptom outcomes. A quarter of our patients had follow-up greater than 5.8 years, permitting a limited evaluation of long term toxicity and symptom outcomes.